Speculations on the growth of ethnobotanical nomenclature

Lang. Soc. i , 51-86. Printed in Great Britain

Speculations on the growth of ethnobotanicalnomenclature1

BRENT BERLIN

University of California, Berkeley

'It is impossible that men, even the most primeval and unlettered, manage their affairswith various denizens of the plant world without classifying them. Names of plants,generic and specific, and also other names more comprehensive, are part of the vernacularof every tribe of the uncivilized, as well as that of every rural province within the bounds ofcivilization to-day. The very names attest the fact of classification; for no name is that ofan individual plant. It is that of a group of plants, always; a group specific, generic or morecomprehensive than either.'

EDWARD LEE GREENE (1909:39)

ABSTRACT

Assuming the ethnobiological classification evolves as a reflection ofcultural development, data are presented which suggest an orderly andpredictable temporal appearance of ethnobotanical nomenclatural categories.A general correspondence is seen to exist between the number of categoriesencoded at any point in time in a particular language's history and degree ofsociocultural development. The principles of lexical marking are applied toethnobiological nomenclature as a means of inferring relative age of thecorresponding categories. (Ethnoscience, primitive classification, languageuniversals, cultural evolution.)

INTRODUCTION

A general observation about the vocabularies of most languages is that they tend

[1] An earlier draft of this paper was prepared with the fellowship support of the Centerfor Advanced Study in the Behavioral Sciences and the National Science Foundation,Grant GS-2280. Essentially this version has been distributed as Working Paper No. 39,Language—Behavior Research Laboratory (March 1971). Many individuals have pro-vided encouragement in developing these ideas further. I am especially appreciativeof the continued collaboration of Paul Kay and William H. Geoghegan of the Language-Behavior Research Laboratory. As much of the data relevant to issues discussed hereare unpublished, I am grateful for ethnographic information provided by colleaguesfrom their own field notes. I would like to thank the following persons for their criticism,data, and helpful comments: Barry Alpher, Eugene Anderson, Robert Austerlitz,Donald Bahr, Keith Basso, Katherine Branstetter, Dennis E. Breedlove, Jan Brukman,Ralph H. Bulmer, Robbins Burling, Wallace Chafe, Harold C. Conklin, LincolnConstance, Christopher Day, Barbara Demory, Robert M. W. Dixon, Mary LeCronFoster, Catherine S. Fowler, Charles O. Frake, Paul Friedrich, William H. Geoghegan,Robert F. Heizer, Nicholas A. Hopkins, Eugene Hunn, Dell Hymes, Paul Kay, RobertM. Laughlin, Yakov Malkiel, Robert McC. Adams, David Price, Robert Randall,Peter H. Raven, Michelle Rosaldo, David Schneider, Brian Stross, Oswald Werner,and Michael Wilson.

51

LANGUAGE IN SOCIETY

to increase in size over time. This pattern is more readily seen when one thinksin broad evolutionary terms, comparing the development of the rudimentarylexicon of early Homo sapiens to the complex vocabularies of modern man. Weknow almost nothing about the causal mechanisms involved in this lexicalexpansion but most anthropologists and linguists would not dispute that it mostlikely mirrors general cultural evolution.

That languages increase the size of their vocabularies through time is, ofcourse, a trivial observation. The study of lexical growth becomes a topic ofrelevance only if it is possible to point out regularities that allow for usefulgeneralizations and predictions about the broader problem of linguistic evolution.

In this essay, I hope to focus on the development of one area of vocabularycommon to most, if not all, languages - names for categories of plants. I willassume that man's vocabulary for kinds of plants has developed over time, anassumption that allows one to ask at least the following questions: Can oneobserve regularities in the ethnobotanical lexicons of past and present day lan-guages that allow one to make plausible inferences as to the major patterns ofnomenclatural growth? If such patterns can be described, are they related toother aspects of man's sociocultural development? And, finally, do such regular-ities appear to have productive implications for the evolution of vocabularygenerally?

I would like to propose in this openly speculative paper that one can recognizesome rather general patterns in the development of ethnobotanical nomenclaturethat are not necessarily self evident. I can not claim that the ideas presented hereare based on well-documented empirical studies. Many of the examples I cite asevidence are drawn from incomplete descriptions. Nonetheless, as further in-formation becomes available, I have become more strongly convinced that thenomenclatural principles sketched here have widespread applicability. I alsobelieve that their recognition may serve as a useful starting point for research onthe evolution of other lexical domains, specifically ethnozoological nomenclature,and, more broadly, on vocabulary generally.

THE SIX UNIVERSAL CATEGORIES OF ETHNOBOTANICAL

NOMENCLATURE

It now appears that the ethnobotanical lexicons of all languages can ultimately bedescribed with the recognition of six major ethnobotanical categories. These sixbasic categories will be labeled as follows: (i) generic, (2) specific, (3) major lifeform, (4) varietal, (5) intermediate, (6) unique beginner.

The names of plant taxa occurring as members of these categories will be,accordingly, generic names, specific names, major life-form names, and so on.

It is furthermore suggested that in the life histories of individual languages,the encoding of each of these nomenclatural categories occurs in a relatively

52

THE GROWTH OF ETHNOBOTANICAL NOMENCLATURE

fixed order. Generic names are fundamental and will appear first. These will befollowed in time by major life-form, names and specific names. At yet a laterperiod, intermediate taxa and varietal taxa will be labeled. Finally, the lastcategory to be lexically designated in the development of any ethnobotanicallexicon will be the unique beginner. The suggested sequence can be seen diagram-matically as follows:

{life formal Cintermediate"! . ,specific / - \varietal / " > U n i q U e b e S l n n e r

Several clarifications of the above sequence should be noted. The first is toindicate that each of the nomenclatural categories, with the exception of theunique beginner, is theoretically an open class. Thus, given the appearance of thegeneric category a language may continue to encode generic taxa throughout itshistory. The same applies to the specific category. The major life form, inter-mediate and varietal categories are likely to be few membered classes whencompared with the generic and specific. This observation is probably a reflectionof the nature of the recognizable discontinuities of the plant world as will beseen in more detail in the following sections.

Secondly, it should also be observed that no temporal ordering is implied forsome categories. Thus, no claim is made as to the priority, in time, of specificnames over major life-form names. On the other hand, a claim is made that alanguage must have encoded at least one major life-form name and one specificname before the appearance of intermediate and varietal named taxa.

As well as noting a general progression in terms of an increase in the numberof ethnobotanical nomenclatural categories in a particular language's history,one may also describe a regular sequence of lexical development for memberswithin each category. Thus, given the appearance of the specific category, onemay observe the further linguistic development of specific names from lexicallyunmarked to lexically marked expressions. The same observation holds for eachcategory. This general feature suggests that languages may not only be rated interms of the number of ethnobotanical nomenclatural categories encoded butcan be ranked as well in terms of the extent to which members of particularcategories have passed from an unmarked to a marked status.

THE REALITY OF NATURAL GROUPINGS OF ORGANISMS

Man is by nature a classifying animal and nowhere is this fact exemplified moreclearly than in his classification of the biological universe. But unlike the some-times capricious and apparently arbitrary classification of certain social phen-omena, all men appear to be constrained in their conceptualization of the world ofplants and animals. It now seems clear that certain naturally occurring groupingsof organisms are recognized as discrete classes in societies which maintain a

S3

LANGUAGE IN SOCIETY

direct and intimate contact with nature. While several ethnographers have longassumed this to be true (cf. especially LeVi-Strauss 1966), the recent work ofRalph Bulmer presents the most explicit statement to this effect. In a series ofperceptive papers on the ethnoscience of the Karam of New Guinea (Bulmer1967, 1968, 1970; Bulmer & Tyler 1968), Bulmer convincingly demonstrates thepsychological reality of such natural groupings. Building on assumptionsconcerning the classification of the natural world which he attributes to Levi-Strauss, Bulmer states t ha t ' . . . in any total classification of plants and animalsthere are important lower order categories which are seen as "objective" by theusers of the classification and which are the smallest logically natural unitsdefined by multiple criteria . . .' (Bulmer 1970:1072).

These minimal, naturally occurring units may or may not correspond in aone-to-one fashion to modern biologically defined taxa, although they generallydo (Bulmer & Tyler 1968; Bulmer 1970; Berlin, Breedlove & Raven 1966;Diamond 1966). They are logically comparable, however, in that in numerousinstances such groupings are formed on the basis of '. . . objective regularitiesand discontinuities in nature' (Bulmer 1970:1072).

The essence of Bulmer's generalizations are clearly in accord with recentresearch inTzeltal botanical ethnography (Berlin, Breedlove & Raven, in prepara-tion), as well as with the work of Conklin (1954), LeVi-Strauss (1966) and otherswho have worked closely with ethnobiological materials.

THE PRIMACY OF GENERIC NAMES

In the ethnobiological lexicons of all languages, one is immediately struck by thestructural uniformity of expressions which linguistically characterize man'srecognition of the basic objective discontinuities of his biological world. Theseexpressions are, for the most part, unique 'single-words' that can be said to besemantically unitary and linguistically distinct. Examples of such semanticallyunitary names in English folk biology would be oak, pine and maple. Primaryterms of this sort appear to represent the most commonly designated concepts ofthe botanical world and can be referred to as generic names. There may or may notbe expressions of greater generality (e.g. tree, vine) or specificity (e.g. black oak,sugar maple), a fact which later will be shown to have important evolutionaryimplications.

An explicit recognition of these psychologically basic ethnobotanical genericterms can be traced ultimately to Theophrastus, the father of Western systematicbotany (see Greene 1909). A more recent exploration of the subject has beenprovided by the ethnobotanist, Harley Harris Bartlett, in his important paper'History of the generic concept in botany' (1940). Bartlett, a good field biologistwith considerable ethnobotanical experience with several Malayan tribes, notedthat a well-defined idea of genera could be found in all of these languages. For

54


Bartlett, t he ' . . . concept of genus must be as old as folk science itself (1940:341).He defined the concept as any class ' . . . which is more or less consciously thoughtof as the smallest grouping requiring a distinctive name' (ibid: 356, emphasisadded). While somewhat vague as a definition, Bartlett's idea concerning thefundamental nature of generic taxa and their corresponding distinctive labelsis essentially correct. In fact, generic names can be seen to exhibit a readilyidentifiable linguistic structure which allows, in most cases, for their immediaterecognition (Berlin 1969, Berlin, Breedlove & Raven n.d.; Conklin 1962: 122;Bulmer & Tyler 1968; Friedrich 1970).

The centrality of named generic taxa as 'semantic primitives' (Friedrich1970: 156) in ethnobotanical classification is important for the growth of ethno-botanical nomenclature. The most obvious significance is that generic names arethe first to become encoded in the ethnobotanical lexicons of all languages. Thus,one may postulate a stage in the development of the plant lexicon of any languagewhereby one finds a series of plant classes, each labeled by generic names, whichpartition a portion of a yet unlabeled taxon best glossed as 'plant'. (It will beshown that 'plant' is the last taxon to be uniquely labeled in any plant lexicon.)

It should be reiterated that the partition of the unlabeled category, 'plant', isnot exhaustive at such a stage and that numerous, potentially namable classesremain unlabeled by generic forms and are linguistically ignored. I would alsoemphasize that folk generic taxa are likely to correspond to botanical generaonly in those cases where the scientific classification reflects obvious morpholog-ical characteristics of plant groupings which are readily observable by simplevisual inspection. Thus, some folk generics will match almost perfectly standardbotanical genera, e.g. oaks, pines, etc., while others will be more inclusive, e.g.cacti, ferns, and so on.

Horizontal expansion of generic names

It is supposed, then, that at some point in the development of a language'sethnobotanical nomenclature one finds a single-leveled taxonomy comprisedsolely of generic taxa labeled by generic names. Over time, groupings of organ-isms earlier not recognized linguistically will be named. If one attempts tospeculate on how this hypothetical early plant taxonomy expanded, the mostplausible argument at the moment is that the direction was, at first, horizontal.By horizontal growth I mean the formation of new generic names. The ling-uistic process of analogy, i.e. when some new category is seen to be conceptuallyrelated to an already existing category and named accordingly, is an extremelycommon form of name formation in contemporary languages. One can assumeit to have been productive at an earlier time as well. Heinz Werner has discussedthis process from a psychological viewpoint and has referred to it as 'concretetransposition'. This process '. . . can frequently be observed in every-day speech.It occurs whenever one uses the expression "something like" or the suffix

55

LANGUAGE IN SOCIETY

"-like" or "-ish", for the description of an object. Concrete transposition isat the basis of many creations of words and changes of meanings..." (Werner1954: 204)

Concrete transposition forms an important part of the naming of behavior ofthe Tzeltal and Tzotzil Mayan Indians of Southern Mexico. When presentedwith a plant that is conceived to be 'related' to a known plant class x, the typicalTzeltal informant will respond kol pahaluk sok x, i.e. 'it is likened to/related tox'. The same semantic information is indicated by the Tzotzil expressionk'os x 'like *'. In both languages, the process is a very common one, allowing forthe classification of the vast majority of all plants in the environment, inasmuchas most plants are seen to be related to some named class. An identical situationhas been reported by Bright & Bright (1965) for the Tolowa-speaking SmithRiver Indians of California and by French (i960) for the Sahaptin Indians ofOregon.

The most thorough-going example of this kind of naming that I have found,however, is reported for a group of Arawak speakers of Surinam. Stahel (1944)notes that 'Arawak Indians in Surinam, when they are naming plants andanimals, make liberal use of the suffix BALLI to reduce the number of primary or"generic" names of the hundreds maybe thousands of kinds they have to distin-guish. For this purpose they have still two other words - DJAMARO and OJOTO. Thefirst means the same as BALLI, the second "related to '" (1944 : 268). As an ex-ample, one may note the following set of Arawak plant names, using Stahel'sorthography.

TATABU Diplotropis guianensis Benth.TATABUBALLI Coutarea hexandra K. Schum.TATABU DJAMARO Copaifera epunctata Amsh.TATABU OJOTO Ormosiopsis flava Ducke.

Stahel states that 'All four are high jungle trees. The first, "zwarte kabbes",is a well-known Surinam timber, the others are less important but all resembleTATABU" (ibid: 269).

There is some doubt that most linguistic anthropologists would treat theexamples mentioned above as legitimate plant names. In a real sense, concretetransposition is a method of making new labels by the use of descriptive phrases.The principle, however, is a productive one and it may become so prevalent, inhorizontal expansion of generic names, that descriptive phrases are eventuallyreplaced by genuine lexical expressions.

Such an advance may be illustrated in certain of the Mayan languages wherebyone notes the use of a generic name plus an animal name to refer to a plant classseen to be related to one indicated by the generic name alone. In Tzeltal onefinds numerous sets like the following:

56


c'omate? 'chayote' (Sechium edule (Jacq.) Sw.)c'omate? c"o 'rat's chayote' (Cyclanthera bourgeana Naud. ex Char.)fihn 'corn' (Zea mays L.)?iHm ?ahaw 'snake's corn* (Anthurium spp.)

k'ewes" 'custard apple' {Annona cherimola Miller)k'ewes" mas' 'monkey's custard apple* (A. reticulata L.).

In Yucatec Maya, the process is also typical, as can be seen in this examplefrom Roys (1931: 223):

Cat 'tree cucumber' {Paramentiera edulis DC.)Cat-cuuc 'squirrel's tree cucumber' (P. aculeata HBK.).

An identical process in naming generic classes which are seen to be related insome form or other is also found in Hanunoo (Conklin 1954) and Subanun(Frake, personal communication), both languages of the Philippines, and Nahuatlof Central Mexico (Paso y Trancoso 1886). In English one notes the pairscabbage: skunk cabbage, apple: horse apple, and oak:poison oak. The usage ofthe English adjectives 'false' and 'mock' may lead to the formation of genericnames in an analogous fashion, e.g. lilac:false lilac; cypress:false cypress;orange:mock orange, and so on.

It is important to note that names formed by analogy of the sort just describedare generic forms and their designata are not conceptualized as subordinatetaxa. Thus, mock orange is not a kind of orange, it is simply like orange in certainrespects. This point is illustrated nicely by reference to the writings of Theo-phrastus as discussed by Edward Lee Greene in his little read Landmarks ofbotanical history (1909). Greene notes that about half of Theophrastus' genericnames are complex expressions including a noun and an adjective. Severalappear to be derived from generic names of a single constituent, e.g. Calamos'reed grass' {Arundo spp.) and Calamos Euosomos 'sweet flag' {Acornus calamus).But Greene has no doubts about Theophrastus' classification of Calamos andCalamos Euosomos as distinct genera. He writes:

It is not imaginable that a botanist of Theophrastus' ripe experience and greatattainments should think those large grass-plants and the sweet-flag to be ofthe same genus. Beyond doubt, however, the name Calamos Euosomos didoriginate in the notion that arundo and acornus are next of kin; for, howeverunlike they are as to size, foliage, and other particulars, there is a remarkablyclose similarity in their rootstocks, these being of almost the same size, formand color in the two. The gatherers of roots and herbs, as we know lookedfirst of all to the 'roots' of things, and these were their first criteria of plantrelationships. To these it should be perfectly natural to place the sweet-flagalongside arundo, the true [Calamos] by its closely imitative 'root', and then

57

LANGUAGE IN SOCIETY

on account of the aromatic properties of the root to call the plant [CalamosEuosomos] (Greene 1909: 123).

Finally, it should be observed that the conceptually central name in a pairconsisting of a generic name and one formed by semantic analogy is lexicallyunmarked, i.e. it occurs in an unmodified form, while the noncentral and his-torically secondary expression is distinguished by a modifier of some sort. Aswill be seen a little further on, this process of lexical marking is a most productiveform of name formation in the overall development of ethnobotanical nomencla-ture.

In summary, I have suggested that the original ethnobotanical vocabularyof any language (and, by implication, the vocabulary of early man) is at firstcomprised solely of semantically unitary linguistic expressions which mark thesmallest conceptually relevant groupings of plants in man's environment.These expressions are known as generic names.

At the outset, the expansion of generic names is accomplished via the processof concrete transposition. It is expressed first in the form of descriptive ordescriptive-like phrases or expressions which at a later period in the history of alanguage may be more formally codified by the formation of definite lexicalforms. This form of expansion is to be joined by the processes of generalizationand differentiation as nomenclature develops over time. Concrete transpositionis to remain, however, as a potentially productive process throughout the historyof a language's development.

DIFFERENTIATION AND GENERALIZATION: THE APPEARANCE OF SPECIFICAND MAJOR LIFE-FORM NAMES

I had at one time hoped to show that specific names become encoded in alanguage's ethnobotanical lexicon before the appearance of major life formssuch as 'tree', 'vine', 'grass' and so on. The data that I have been able to gatherat this time do not allow for a definitive answer as to which ethnobotanicalcategory may be prior. I know of no language which lacks at least some specificplant names, although there may have been languages, such as that spoken by theTasmanians, which lacked general life-form terms. The evidence on this point,however, is scant and probably unreliable.

Consequently, a weaker hypothesis is presented here which posits no temporaldistinction as to the rise of specific and supra-generic categories, although furtherresearch may require a modification of this view.

DIFFERENTIATION AND THE FORMATION OF SPECIFIC NAMES

In Bartlett's paper on the genus concept, he noted that: 'With enlarging

58


experience, people make finer distinctions and need different names for newlydistinguished entities which have previously been called by the same originalname. The original name becomes generic in its application; variously qualifiedit provides the basis of specific names' (1940: 349).

I would take issue only with Bartlett's phrase '. . . The original name becomesgeneric in its application . . . " It had originally been generic and, with enlargingexperience, is merely partitioned into sub-classes.

There appears to be some psychological evidence to support such a position.Werner notes ' . . . that the predominant developmental trend is in the direction ofdifferentiation rather than of synthesis. [Likewise], the formation of generalconcepts from specific terms is of lesser importance in non-scientific com-munication though it is rather a characteristic of scientific endeavor. In otherwords, language in every-day life is directed toward the concrete and specificrather than toward the abstract and general. Because of this trend toward theconcrete, semantic generalization develops slowly and by intermediate steps'(Werner 1954: 203).

If differentiation is to occur and be lexically encoded, there appears to be afairly concise way in which one can imagine it happening. First, the division of ageneric name is most probably binary, at least at the outset. This observation isborne out in fact in present day folk taxonomic systems which have been wellstudied. By far the greatest number of contrast sets comprised of specific taxain folk biotaxonomies are comprised exactly of two specific names (Berlin,Breedlove & Raven in press; Conklin 1954: 128).

Type-specific nomenclature

A highly regular labeling process can be described for the encoding of specifictaxa, given the primarily binary partition of a generic taxon. In general, onespecific category, because it is most widespread, larger, best known, or the like,will always be recognized as the typical species of the folk genus. This taxoncan be referred to as the type-specific, the archetype, or the ideal type. 'Type-species' have long been recognized in systematic biology. The notion, codifiedby Linnaeus, can be seen to have its origin in folk biosystematics since earliesttimes and has been reported by many anthropologists working with societiesintimately involved with nature. Dentan, who has worked extensively with theprimitive Semai of Southeast Asia, notes: 'My impression is that the Semai thinkthat some species are more "typical" of their categories than other species are.For example, naga (snakelike dragons) seem to represent the quintessence of"they beneath the earth". Giant monitors, in turn, are the prime representativesof "lizard", and regal python of "snake"' (Dentan 1968: 35).

A strikingly general nomenclatural regularity can also be suggested as regardsthe relationship of the type specific, its contrasting non-typical specific, and thesuperordinate generic name. In the early stages of the division of a generic

59

LANGUAGE IN SOCIETY

taxon into two specific categories, the type-specific will invariably be polysemouswith its superordinate generic in characteristic usage. Stated in other terms, thetype-specific taxon will be lexically unmarked in normal speech, being referredto by the identical linguistic expression as its superordinate generic. Such poly-semous labeling of taxa at differing levels of generalization has been discussed byHymes (1960), Frake (1962) and Conklin (1962) in reference to folk biosyste-matics and by Greenberg (1966) in reference to lexicon in general. Severalexamples will illustrate this nomenclatural principle. As Wyman and Harris havesaid in referring to Navaho ethnobotany, 'The situation is as if in our binomialsystem the generic name were used alone for the best known species of a genus,while binomial terms were used for all other members of the genus' (1941: 120).

Washington Matthews (1886) was the first to recognize explicitly type-specificnomenclature in Navaho. In an especially careful piece of research for its time,Matthews notes that '. . . there are three species of juniper growing in theZuni mountains; each has its own appropriate name, yet the generic name forjuniper . . . appears in all' (1886: 767). Diagrammatically:

kat'Juniperus spp.'

katjfuniperus communis L.

'common juniper'

kat-nee-ay-lijf. virginiana L.'strained juniper'

kat-dil-tah'-liJ. pachyphaloea Torrey

'cracked juniper'

Several other examples found in Matthews show the same principle at work:

tlotdhi

tlotdhi tlotdhi-tsoChenopodium C. album

fremontii 'great tlotdhi'

tsinya tlotdhi tlotdhi-hochiTeloxys cornutum Amaranthus retroflexus'tlotdhi under the 'prickly tlotdhi'

trees'

60


a-xay-in-kliri-i

a-zay-in-klin'-iMahastrum coccineum

'gummy or glutinous medicine'

a-zay-in-klini-tsoSphaeralcea fendleri

'large gummy or glutinous medicine'

tsa -si

tsa'-siYucca baccata

'yucca'

tsa'-si-tsozY. angustifolia'slender yucca'

Classical Nahuatl also exhibits this nomenclatural regularity in that the mostcommon specific form included in a particular generic class is labeled poly-semously by the generic name. Thus, one observes that in the Nahuatl classifica-tion of sedges, tollin included a '. . . type-species [sic] that carried simply thename Tollin and that [also] referred to the sedge family, various other relatedspecies of it having been grouped under the same name, each with a differentdetermination' (Paso y Troncoso 1886: 218).

A final example of polysemous generic and type-specific plant names can beseen again in the early work of Theophrastus, who, as a plant nomenclator, tookgreat pains to preserve the essential structure of the names discovered to be incommon usage in his day. As Greene says, Theophrastus ' . . . left plant nomen-clature as he found it' (1909: 123). Greene captures the essence of the earlybotanist's view aptly: 'The Theophrastan nomenclature of plants is as simplynatural as can be imagined. Not only are monotypic genera called by a singlename; where the species are known to be several, the type-species of the genus -that is, that which is most historic - is without a specific name, at least verycommonly, and only the others have each its specific adjective superadded to thegeneric appellation' (Greene 1909: 120).

The following examples attest to this fact:

[Theophrastus' names]PeucePeuce Idaia

[Modern equivalents]Pinus piceaP. maritima

61

LANGUAGE IN SOCIETY

Peuce conophorosPeuce paralios

MespilosMespilos anthedon

P. pineaP. halepensis

Mespilos cotoneasterCrataegus tominalis

Lexical marking of type-specific namesWhile many languages find no need to mark linguistically the focal type-specificwithin a particular grouping of specific names, a circumstance which I suggest ischaracteristic of the process employed in the earliest specific name formation,situations of social intercourse may arise whereby one must be able to ling-uistically differentiate the type-specific category from its contrasting neighbor(s).The linguistic process by which this contrast comes to be indicated is also quitegeneral. Invariably, the type-specific will be modified with an attributive-likeexpression best glossed as 'genuine', 'real', or 'ideal-type'.

Such a situation is found characteristically in many languages, of whichTzeltal and Hanun6o may be cited as examples.

In Tzeltal, He is the generic name for 'chili pepper' {Capsicum spp.). In mostcontexts, He can be used alone to refer to the most prominent specific class.However, when greater precision of designation is required, the attributivebac'il 'genuine' is readily applied to distinguish this specific class, bac'il ?ic'genuine chili pepper' from its contrasting coordinate specific classes. Thus, onefinds:

(bac'il) ?ic'genuine chilipepper'

cawcaw ?ic

'round chilipepper'

cahnut ?ic'chicken feceschili pepper'

ton He

'stone chili pepper'

An identical nomenclatural principle has been described for Hanun6o.Conklin, in an as yet unpublished ethnobotany of this people, notes that ' . . . ashared term [i.e. a generic plant name partitioned by two or more specificnames] when not followed by an attribute, may be read as that term plus ?urunan"real". The resulting name is a preferred synonym, required where the desig-nated plant name is distinguished from others in the same set' (Conklin 1954:259). An example can be seen in the classification of Job's tears.

6a


palyas'Job's tears'

palyas (uruyari)'genuine Job's tear'

palyas hintahay'smaller Job's tear'

While the type-specific remains unmarked, or marked only in contexts whereambiguity might arise, the non-typical specific(s) will obligatorily be marked.The linguistic structure of such non-typical (or secondary) specifics appears tobe of a specifiable form. Such names will be comprised of the generic names (inwhich they are included) plus a modifying attributive-like expression. In all cases,such non-typical specific expressions will be binomial in structure.

The feature(s) focused on by the modifying attributive-like expression isgenerally some obvious perceptual dimension such as color, size, growth habit,habitat, or the like. Such a situation may be diagrammatically indicated in thefollowing hypothetical specific contrast sets.

GENERICS

SPECIFICS (genuine) x red x (genuine) y small y (genuine) z water-place z

Fossilization of the type-specific attributiveIn some languages, for example, especially the Mayan language, Tzeltal, 'patternpressure' appears to be working so as to make the presence of the type-specificattributive (i.e. 'genuine') obligatory or independent of context. As an example,the type-specific bac'il ?alcas 'genuine orange' is almost universally the preferredusage (vs. the simple unmarked ?alca£) in contexts where it contrasts with formssuch as pahal ?alca$" 'sour orange' ?elemones ?alcai 'lemon orange', etc. Thistendency, I think, represents a later development which follows logically fromthe prior, unmarked usage.

An even further logical sequence can be seen at work in Tzeltal which may ormay not have general validity as a subsequent development in all systems. Infew membered contrast sets having two or three members one notes the tend-ency for the type-specific to assume a value on the semantic dimension indicatedin the name(s) of its contrasting member(s). Thus, while a contrast set mighthave at one time included members labeled as:

LANGUAGE IN SOCIETY

X

bacHl x 'genuine .v' cahal x 'red .v'

where the semantic dimension of color appears indicated in the marked non-typical specific, it now becomes habitually labeled as:

sakil x 'white x' cahal x 'red .v'

As concerns specific examples where the dimension of color is involved, Ithink it of no mean theoretical significance that the attributive replacing 'genuine'in the type-specific member is sakil 'white', a rather 'typical color' in flowerpigmentation (flower color being a major semantic dimension used to differen-tiate many closely related specifics) and almost neutral, as it were, in terms of itsmarking potential.

The development of lexical markings in specific nomenclature

I now want to summarize the theoretical developmental sequence for the lexicalmarking of specific plant names.

(a) First, the generic taxon is partitioned into a type-specific and one (or more)non-typical specifiers). The type-specific is lexically unmarked and polysemoiiswith its superordinate generic. The non-typical specific(s) is lexically marked, afeature leading to a binomial expression. Diagrammatically:

GENERIC

SPECIFICS x a x (where a — attributive)

(b) Over time, the type-specific must be linguistically distinguished in certaincontexts of semantic contrast from the non-typical specific(s). An optional'type-marking' attributive is applied to the type-specific. Invariably it will bestbe glossed as 'genuine', 'real', or 'most typical'. Diagrammatically:

64


GENERIC .V

SPECIFICS . (genuine) .v a x

(c) In the penultimate stage, time, usage and binomial pattern-pressure ofspecific nomenclature will force the type-marking attributive, 'genuine', tobecome obligatory. Once obligatory, its semantic marking function is radicallyreduced and ultimately it becomes completely neutralized. Diagrammatically:

GENERIC

SPECIFICS genuine x a x

(d) In the ultimate stage, pattern pressure, will force the neutralized type-marking attributive to be replaced by an attributive from the same semanticdimension as the attributive indicated in the contrasting non-typical specific(s).Thus, a set of forms that were formally:

genuine x

will become:..

red .v genuine y small y

white x red x large y small y

THE RISE OF MAJOR LIFE-FORM NAMES

While generic and specific names account for the vast majority of names forplants in natural ethnobotanical vocabularies, a smaller number of forms occur

65

LANGUAGE IN SOCIETY

which are of greater inclusiveness than generics and specifics. These expressionsmay be referred to as life form names. At some point in man's classification of theplant world, the concrete differences marked by generic names could be dis-pensed with in some contexts and higher order abstract names were developedwhich marked such major life forms as 'tree', 'vine', 'herb' and so on. This is notto say that such categories had not been conceptually recognized by man sinceearliest times. I find it difficult to conceive that such was not in fact the case. It islikely, however, that these higher order categories lacked simple, monolexemicdesignations. The development of life-form names is most certainly subsequentto the appearance of generic names in the evolution of ethnobotanical nomen-clature.

While very little is known concerning the processes involved in the formationof life form names, it is clear that in many languages the labels for life formcategories are drawn directly from the existing inventory of generic names.Furthermore, the labeling of major life form taxa can be seen to follow identicalmarking principles as those described for generic and specific names. Applyingthe principles which we have seen to be at work in the case of generics and specif-ics, one can make fairly good guesses as to which names get elevated to majorclass status: precisely those generic names which, because of their distributionand cultural importance, are most salient culturally. In many of these instances,the life form name and a subordinate generic are polysemous.

Buck's massive compilation of synonyms for the major Indo-Europeanlanguages provides data on the rise of the major life form name "tree" that bearson this hypothesis: 'A widespread group of words for "tree", many of themmeaning also "wood", go back to an IE word which probably denoted a par-ticular kind of tree, namely the oak' (Buck 1949: 48). And again: 'Noteworthy isthe primacy of the oak, as shown in mythology and in the recurring use of "oak"as the tree par excellence, for "tree"' {ibid.:528). Finally, Buck notes thereconstruction "Oak". 1. IE* derwo- dru-, etc. in words for "oak", and for"tree", "wood", the former, specific use being probably the earlier' {ibid.).

The most recent and authoritative statement on the proto-Indo-Europeanarboreal system is that of Paul Friedrich (1970). Friedrich rejects the generallyaccepted alternative hypothesis that *derwo originally meant 'tree', rather than'oak'. His conclusions are especially interesting:

. . . it seems probable that the primitive, arboreally oriented PIE distinguishedseveral species of oak by distinct morphs, and that *ayg-, *perkw- and *dorw-served in this way. As the oak and mixed-oak forests were reduced and con-tracted, and as the speakers of the PIE dialects migrated into their newhomelands - two simultaneous processes during the third and second mil-lennia - the denotations of the *dorw- reflexes shifted to 'wood, tree, hardness,'and yet other referents; this would hold especially for the shift to 'fir, tar,

66


pinewood' and the like in the Baltic and North Germanic dialects, since thespeakers are thought to have migrated into northern coniferous zones duringthe centuries when the oaks were receding. It is also quite possible that even inPIE times the main name for the oak - a sort of Urbaum - was occasionallyor dialectically applied to 'tree' in general. Within pre-Homeric Greek 5pusand 6pu6s could denote either 'oak' or 'tree' with disambiguation throughsocial or literary context. By Classical Greek times the meaning had narrowedto the original PIE 'tree'. In more recent centuries the identical process hasbeen documented in Germanic, where eik shifted from 'oak' to 'tree' inIcelandic - oaks being virtually absent in that country (Friedrich, 1970: 146).

While the polysemous origins of life form names in nomenclaturally advancedlanguages have become obscure, their etymologies determinable only by histor-ical reconstruction, such is not the case with numerous less advanced societiesyet in the early stages of their nomenclatural life-histories. In fact, some datasuggest that one is observing in some languages the actual accension of somesupra-generic taxa, suggesting that such languages have only recently movedfrom the prior stage of ethnobotanical nomenclature of generic names only.Furthermore, it is not surprising to find polysemous generic and life form plantnames in languages spoken by societies which are rather simple in their cultural-technological development.

The best reported cases now available to me are found in the Great Basin andthe Southwestern areas of the United States, though there are doubtless otherexamples in other parts of the world. Trager (apparently unaware of an earlierpaper by Albert Gatschet [1899] who reports the same data) noted that in severalSouthwestern Indian groups, the word for 'tree' was polysemous with the wordfor 'cottonwood', the only deciduous tree which grows with abundance outsidethe major forests. For example, in Taos Pueblo '. . . the ordinary word for"tree" . . . tuldna is also the word for "cottonwood"' (Trager 1939:117). Further-more, 'In Isleta and Sandia (southern Tiwa dialects not very different from thenorthern Taos and Picuris) the word tula means "tree, cottonwood" . . .'(ibid.). Finally, in Hopi, '. . . we find the same word, sohdvi, used for both'(ibid.).

A more recent study by Fowler and Leland on Northern Paiute ethno-systematics verifies Trager's observations for this Great Basin people. Theystate t h a t ' . . . the terms for cottonwood tree [sigdbi] can include the willow treeat one level and can also be used in popular speech for any deciduous tree'(Fowler & Leland 1967: 387).

In a paper recently brought to my attention by Barbara Demory, Almstedt(1968) reports generic name - life form name polysemy for Diegueno, a Yuman-speaking group of Southern California. Here, however, the term for 'tree',isnyaaw, is polysemous with another ecologically important tree, the California

67

LANGUAGE IN SOCIETY

live oak, Quercus agrifolia. This species of oak has the widest distribution and isthe most generally available source of edible acorns in the area inhabited bythese people. For Almstedt,"... it seems logical that this name isnyaaw should beused for tree when the need arose' (1968: 13).

Demory (in press) surveyed additional languages in the Hokan family andfound numerous other examples of life form-generic polysemy. In each case,the generic of major cultural significance in that particular geographic areaappears to have assumed life form status. Thus in Karok, HpahA 'juniper, tree';Achumawi, aswo 'sugar pine, tree'; Atsugewi, ajwi 'kind of pine, tree'; YanabaacuVi 'broad leaf maple, tree'; Salinan, hat' 'oak, tree'; Chumash, ku-wu 'liveoak, tree'.

Earlier work by Bright & Bright (1965: 253-4) reports the term tepo aspolysemously meaning both 'fir' and 'tree' and Gatschet notes Klamath k'osh asboth 'pine' and 'tree' (1890: 1, 146).

A final example of this nomenclatural regularity is found, in Western Apacheas described by Keith Basso. It is made even more important because Basso'sdata bear both on the rise of major life-form names as well as the formation ofspecific names. Basso writes:

'The situation among the Western Apaches is much the same as that youdescribe for the Great Basin and other portions of the Southwest. The termfor cottonwood (t'tis) is also used for 'tree'; in addition, however - and this iswhat makes it interesting -t'tis may also designate a 'real cottonwood' - namely,those which are tall, heavily foliaged, and situated near the banks of flowingstreams and creeks. A few cottonwoods, much more stunted and less greengrow in dry washes and arroyos. These are called t'tis da'iskqq ('cottonwoodsunderfed'). T'tis in the sense of 'genuine cottonwood' is sometimes labeledt'tis da'bnhii 'cottonwood true'/'cottonwood correct' (Basso, personal com-munication).

Basso diagrams the taxonomic structure of these lexical items as follows:

t'tis 'tree'

t'tis 'cottonwood' [other trees]

t'tis da'bnhii t'iis daiskdd'genuine cottonwood' 'underfed cottonwood'

68


A linguistic expression does not become a major class form overnight, how-ever. Presumably, the process is a relatively slow one. Data from Shoshone,another language of the Great Basin, may be relevant at this point. Wick Miller,who has been involved in an extensive study of the Shoshone, notes the commonuse of sohopi for tree, cottonwood, willow, aspen, stick and log, but is uncertainas to whether the term can be applied to any tree, e.g. 'What I do not know, andwhat would be of particular importance to you, is whether or not these terms[sohopi] can be used generically [i.e. inclusively] that is, can I say, for example,"an oak tree is a sohopi" ("cottonwood tree")?' (Miller, personal communica-tion).

Fowler (personal communication), on further consideration, also questions theuse of Northern Paiute sindbi for all deciduous trees, suggesting that the termmight be restricted to aspen, cottonwood and willow. Ash, alder and mountainmahogony, potential candidates for sindbi, grow outside the Northern Paiute range.

Miller's and Fowler's comments bring up an interesting point as concerns theassumption of supra-generic status by a particular generic taxon. We have noreason to assume that a recently elevated supra-generic should have, at theoutset, an extension radically different from that of the generic from which it hasarisen. Thus, a term which may eventually come to refer to all 'hard, single-stemmed, erect plants attaining a specified height' (i.e. 'trees'), might originallybe restricted to a small sub-set of 'trees'. In the case of the Shoshone termsohopi, it might be restricted to 'cottonwood-like deciduous trees' (e.g. cotton-wood, willow, aspen, etc.) and later come to refer to deciduous trees generally.An even later development might force the extension of the general term toinclude the obviously aberrant conifers, e.g. pines, firs, etc. This surely musthave been the case with the general use of Indo-European *derwo from 'oak' to'tree' in general. A comparable situation is illustrated in Tzeltal where 'tree' isseen to include only truly woody stemmed organisms attaining a specified height,excluding' such 'obvious' trees (from the folk western point of view) as palms,these latter forms being considered as unique generic classes.

The ultimate extension of the major class supra-generic 'tree' can perhaps bestbe illustrated in modern American English folk botany where any single-stem-med upright plant with leaves at the top is admitted to the class, allowing for thebotanically unlikely assortment of such diverse organisms as oaks, pines, palms,banana (trees) and even certain tree-like bamboos.

The marking of 'type generics'

The examples I have thus far cited have all referred to languages where theculturally central generic term which gives rise to major life-form names remainsunmarked, in common usage, being polysemous with its superordinate supra-generic. Reviewing examples of generic-specific polysemy, might one not alsoexpect to find systems whereby the generic name becomes marked, at least in

69

LANGUAGE IN SOCIETY

certain contexts? And might not these instances be seen as a subsequent develop-ment, just as was the case in generic-specific polysemy? In fact, such examplesare found, though in the fossilized stage of development, in Tzeltal and Tzotzil,although there is also evidence of a similar situation in Kiowa Apache of Okla-homa.

The marking processes distinguishing what now might be called the 'type-generic' are identical to those described earlier for type-specifics. Thus, ifmodified, the polysemous form comes to be qualified linguistically with, anexpression best glossed as 'genuine'.

To illustrate, in Kiowa Apache, the term for 'tree' is ?ddw while that for'cottonwood', the most prominent deciduous form, is a-hi, a form literallytranslated as 'tree-real', i.e. 'genuine tree' (see Trager 1939: 118). For somespeakers of Tzeltal, the term bac'il ?ak' 'genuine vine' (< ?ak' 'vine') occurs as ageneric for the most important vine utilized in house construction binding,Smilax subpubescens A.DC. One also finds, for some speakers, the genericbac'il ?ak 'genuine grass' (<?ak 'grass') for the most common and importantgrass employed as a major thatching in house roofing, Muhlenbergia macroura(HBK). Hitchc. Finally, in Zinacantan Tzotzil, the most important tree, both inhouse construction and as a firewood, is bac'i-te? 'genuine tree' (<te ? 'tree'),a form which refers to the prominent oak species of the area, Quercus peduncularisNee.

These terms in the Mayan languages just cited are not metaphorical orsynonymous expressions for the informants who have them as terms in theirrespective nomenclatural systems, although synonymous expressions do exist.The fact that the attributive constituent bac'il 'genuine', has become frozen ineach case, its presence being obligatory, is quite analogous to the situationswhereby the type-marking attributive fossilizes and becomes obligatory inspecific names.

Finally, Gatschet (1899) reports a similar body of data from Nipissing, adialect of Ojibwa. Here andak means 'evergreen tree' and inin andak 'realevergreen tree' or 'pine'. Likewise, atik 'deciduous tree' includes the genericname inin dtik 'real deciduous tree', i.e. 'maple'. One can see how a culturallycentral generic name in each case has become obligatorily marked linguisticallyby the form 'genuine'.

Summary of the development of life-form names

The theoretical development sequence for the appearance of major life formtaxa may now be summarized as a series of at least four steps.

(a) At the outset, the newly encoded life form category is labeled polysem-ously with the most common generic form from which it was derived. Itsextension is, at first, rather limited, perhaps only to those generics which areseen to be quite similar to the type-generic. Diagrammatically:

70


LIFE FORM X

GENERICS x [type-generic]

(b) Over time, the type-generic comes to be optionally marked, linguisticallywith the type-marking attributive, 'genuine'.

LIFE FORM

GENERICS (genuine) x

(c) Further development leads to the fossilization (neutralization) of the type-marking attributive, 'genuine', and it becomes obligatory. Meanwhile, the lifeform is expanding its referrential extension, including taxa originally excluded.

LIFE FORM

GENERICS genuine

(d) Final stages of growth are indicated when polysemy is totally obscured inthe current forms, the original expressions having been replaced or otherwisechanged.

LIFE FORM

GENERICS genuine x

THE APPEARANCE OF VARIETAL NAMES

It would appear that the naming of varietal taxa will, in general, follow theappearance of the major life-form taxa. This is not a contradiction, as it might atfirst seem, of the tendency for abstract terms to occur later in a language's

71

LANGUAGE IN SOCIETY

history than highly specific ones. It is a reflection of the greater control over theprocesses of domestication that man has acquired only through long millenia oftrial and error methods. All of the information available to me at the momentshows that legitimate varietal names occur almost exclusively in the classificationof important cultivars. I would imagine that the same holds true for animals.The control over nature that is required in selecting and maintaining a par-ticular race of corn, beans, rice, chili-pepper, squash or what have you, mor-phologically distinctive enough to merit habitual lexical designation, can beaccomplished only by relatively advanced horticulturalists. Accordingly, oneshould not expect to find varietal ethnobotanical nomenclature except in thelanguages of societies which practice rather refined methods of cultivation. Evenin these languages, varietal names will be restricted to highly important groups ofcultivated plants.

Given the appearance of varietal taxa, however, it appears relatively simple tospecify the linguistic structure of names used to refer to such taxa. In all cases,the specific name (logically of the form attributive, generic, order being irrelevant)will be formed by the addition of another attributive. And, as one might expect,the processes of lexical marking described earlier apply equally well in varietalname formation.

The general principles seen to be at work can be illustrated firstly in Tzeltal.In this language, the generic taxon lo?bal 'banana' includes at least twelve distinctspecific classes, e.g. bac'il lo?bal 'genuine banana', sakil lo?bal 'white banana',sera lo?bal 'wax banana', cahal lo?bal 'red banana', etc. One of these specificforms includes two varietals labeled in the expected manner, i.e. the type-varietaloccurring unmarked, in typical usage, but marked with the type-markingattributive, 'genuine', if ambiguity occurs, and the non-typical varietal beingobligatorily marked. Thus one notes the forms:

(bac'il) cdhal lo?bal '(genuine) red banana'sakil cahal lo?bal 'white red banana'

It is rare that varietal names include more than two modifying expressions,i.e. indicate classes of greater specificity than 'first-order' varietals. In Hanun6o,of the 1094 terminal taxa that are specific or varietal names, 961 are specifictaxa and 97 are varietals. Of these 97 varietals, 87 are first-order forms, i.e.marked only by two attributives. Only two generic names include 'sub-varietals'comprised of more than two attributives and these occur in the highly importantcultivars, corn and chili-peppers (Conklin 1954).

In actual speech, it is also rare that a varietal be referred to by its 'full name'.There is a strong tendency for such forms to participate in what Conklin hascalled 'abbreviation' (Conklin 1962: 122), i.e. when a part of the name may beused to stand for the varietal class as a whole. In such cases, abbreviation willlead the primary, i.e. specific, attributive to function as a head of the expression,

72


the resulting form in most cases being, then, binomial in form. As an examplenote the English varietals:

butter lima(s) beanbaby lima(s) bean

where butter lima(s) and baby lima(s) may occur as complete expressions. Thesame can be noted in Tzeltal, also for beans, where one notes the forms:

cahal slumil cenek' 'red ground beans'?ihk'al slumil cenek 'black ground beans'

where cahal Humil 'red ground [ones]' and ?ink'al slumil 'black ground [ones]'can occur alone.

THE PROBLEM OF INTERMEDIATE TAXA

In examining the ethnobiological lexicon of numerous languages, I havenoticeda strong tendency for the hierarchical depth of biotaxonomies to be uniformlyshallow. Superordinate taxa of greater inclusiveness than the folk genus, thelife form names, are invariably few in number and are inclusive of the majorityof all named taxa. Subordinate taxa of lesser inclusiveness than folk genera,i.e. specific names, are likewise few in number and occur predominantly in thosetaxa with critical cultural importance (e.g. cultivated plants or domesticatedanimals). One may generalize and claim that most folk biotaxonomies tire com-prised primarily of named generic, major life form, and specific taxa, with genericclasses being by far the most numerous and psychologically significant.

Covert 'mid-level' categories of greater inclusiveness than folk generic cate-gories but not yet life-form categories may be seen to exist in many taxonomiesand their recognition is of crucial importance to a full understanding of thecomplete classificatory structure (see Berlin, Breedlove & Raven 1968). However,the fact that these mid-level categories have not been labeled suggests that theneed to distinguish such classes is as yet relatively unimportant in most culturalcontexts.

Nonetheless, the question remains: Why are named intermediate taxa almosttotally absent in natural ethnobiological taxonomies? The conclusion that I havetentatively come to is that such taxa are rare because they are basically unstablecategories, a point which will be developed below.

How are named intermediate taxa likely to arise? At one point, it was suggestedthat the already present covert categories of this taxonomic rank would be themost probable candidates for labeling (see Berlin, Breedlove & Raven 1968: 297).As research continues, this hypothesis appears not to be verified in fact. Whathas been discovered, however, is that named taxa of less inclusiveness thanmajor life forms yet more inclusive than folk generics appear primarily as a

73

LANGUAGE IN SOCIETY

response to situations whereby native polytypic generics must be distinguishedfrom newly encountered generics (see now Berlin, Breedlove & Raven, n.d., fora detailed discussion).

Thus far, two distinct processes - or better said, paths - have been observedthat account for the rise of named intermediate taxa. The first occurs in culturecontact situations where certain introduced organisms must be incorporated intothe native taxonomy. If the introduced plants are conceived to be similar - inthe native view of the world - to a named polytypic native generic class and yetnot similar enough to be included as a specific of that generic, a named higher ordertaxon will arise which includes both the native and introduced forms.

The second process, not as clearly understood as the first, occurs when somespecific taxa become 'conceptually' distinct from their neighboring specifictaxa. When this occurs, the conceptually distinctive taxon will assume the status ofa generic, will cease to be labeled by a binomial expression, and in so doing, will forcethe original generic to assume a superordinate taxonomic status.

The first process can be illustrated with examples from Tzeltal. At the time ofthe Hispanic Conquest, the highland Mayan groups were introduced to twosimilar and yet quite distinct grain crops, wheat and sorghum. These grain-bearing crops were considered to be similar by the Tzeltal population to theirown polytypic generic class of native corn, & . Logically enough, the twointroduced classes were linguistically designated as kaslan ?i!!im 'Castillian corn',i.e. 'wheat' and moro ?isim 'Moor's corn', i.e. 'sorghum'. Their conceptual affilia-tion with corn is verified in that both names occur as responses to the query,bitiksbilhuhuten?isim'What are the names of each kind of corn?' Further question-ing, however, clearly demonstrates that these two introduced plants are not kindsof genuine corn, or, as the Tzeltal would say, not bac'il ?isim.

The new taxonomic structure, then, is seen as one where a superordinateclass of greater inclusiveness than those which have been considered as genericgroupings, has arisen. Diagrammatically:

?isim 'grains'

{bac'il) ?isim 'corn' kaslan ?isim 'wheat'(i.e. 'Castillian corn')

moro ?isim 'sorghum'(i.e. 'Moor's corn')

k'anal Hsim'yellow corn'

sakil ?isim cahal ?isim ?ihk'al ?isi?n pintu ?isitn'white corn' 'red corn' 'black corn' 'multi-colored corn'

74


The newly formed superordinate taxon, ?i$im, may best be glossed as 'grains',at least for the moment. And, in precise accord with the processes of lexicalmarking described earlier, the type-generic 'corn' is marked optionally by theattributive bac'il 'genuine' distinguishing it, in situations of ambiguity from'wheat' and 'sorghum'. Otherwise, it is polysemous with the newly formedsuperordinate intermediate taxon, 'grain'.

Nomenclaturally, the recently introduced generic names are binomial instructure and in this respect do not conform to the otherwise linguisticallyunitary structure of other generic names. Time and usage, however, will tend toneutralize the marking properties of the attributive forms kailan and moro andthe expressions will come to be conceived of as single, semantic units. Geogheganhas aptly characterized the process as follows:

Either because of frequency of perception, need to communicate or what-have-you, a single pattern can be established . . . i.e., rather than identification pro-ceeding from the use of two patterns, as in the first stage, a single unsegmentedpattern recognition routine comes into use. At this point, the [complex]nature of the coding no longer has support (since only one feature rather thantwo is being used), and the complex term will have a tendency to decay,[becoming] a lexeme with a unitary cognitive representation (Geoghegan,personal communication).

A strikingly similar situation to that described for the Tzeltal data can be seenin the classificatory treatment of the introduced New World sweet potato(Ipotnea batatas), among the primitive peoples residing in the vicinity of MtHagen in the Central New Guinea Highlands. The Mt Hagen material illustrates,furthermore, the interplay of lexical marking and cultural significance in aninteresting and important way.

In Hagen ethnoscience, oka refers to the sweet potato, / . batatas. Oka mapumb,a contrasting generic, refers to the indigenous edible tuber, Pueraria lobata.A third name, oka koeka, refers to a wild, inedible tuberous vine and oka kombklato a wild tuberless vine with leaves similar to oka.

Marilyn Strathern reports that all of the four above names are consideredconceptually similar to one another in terms of a variety of characters. On theother hand, it is clear from Strathem's paper that each form refers to a distinctivegeneric class. Thus, oka mapumb, 'Pueraria' is not a kind of oka.

Oka mapumb (Pueraria) may be contrasted with oka ingk (true [ingk] oka,i.e. sweet potato) or with oka alone, which, when unqualified, always refersto sweet potato [sic]. Only if modified by mapumb, koeka, etc., does oka meansomething other than sweet potato. Conversely, Pueraria can only be referredto by employing a special suffix such as mapumb; it is never just oka (Strathern1969: 193).

75

LANGUAGE IN SOCIETY

Furthermore, additional evidence shows that oka is partitioned into variousspecific names, none of which include these forms oka mapumb, oka kombkla, etc.

In Strathern's words:

Oka (sweet potato) may be divided into numerous secondary taxa. Whencollections of these names were made, only once did mapumb enter any list.In. all the other cases the secondary taxa referred to divisions of oka = sweetpotato, and excluded any mention of oka mapumb. Mapumb is thus not seen as anamed type of sweet potato on a par with the other varieties (konome, pora,etc.) (Strathern 1969: 193). • .

The linguistic and ethnographic evidence suggests that one may characterizethe Hageners classification of the plants involved as follows:

(oka) 'sweet potato-like vine'

oka (ingk) , oka mapumb oka koeka oka kombkla'(true) sweet potato' 'Pueraria lobata' 'wild, inedible tuber' 'wild tuberless

vine'

oka konome ' oka pora etc. . . . etc.'specific classes of sweet potato'

The Mt Hagen example, then, can be said to be strikingly parallel to theTzeltal treatment of grains. Here one sees the possibly covert recognition of anintermediate taxon, oka 'sweet potato-like vines' that includes not only sweetpotato but as well Pueraria and related vines. Unlike the Tzeltal case, however,the introduced generic, oka, has come to assume unmarked status, being optional-ly marked only in contexts of ambiguity. While the details of this developmentare unclear, one may make several inferences which are fairly well supported onlinguistic grounds.

It is likely that at one point, before the introduction of the sweet potato, oka,referred, in its unmarked form, to Pueraria lobata. The terms oka koeka 'wildinedible tuber' and oka kombkla can be seen simply as examples of generic nameextension formed on the basis of analogy as discussed in p. 55 above.

With the appearance of sweet potato, which is morphologically similar toP. lobata, one can surmise a period whereby the former class was linguistically

76


referred to as oka x, where * must have represented some unknown qualifyingexpression.

Through time, the cultural importance of sweet potato increased dramatically,eventually exceeding that of P. lobata. At some point in this process, oka cameto refer, in its unmarked form, to sweet potato and P. lobata became obligatorilyindicated by the complex expression oka mapumb.

In fact, a process identical to the. one just supposed can be documented withsome accuracy for certain Tzeltal animal names (see p. 80 below).

The second process whereby superordinate intermediate taxa are named doesnot necessarily result from the introduction of new organisms. This situationoccurs when a native specific assumes the status of a generic category. There may bea period in the process when the distinctive (conceptually) specific taxon islabeled by a unitary lexeme and not the standard binomial expression char-acteristic of most specific taxa.

The process can be illustrated by an example from Tzeltal and concerns theclassification of oaks. For most informants, the generic hihte? 'oak' includes fourspecific taxa, ca?pat hihte? 'excrement barked oak', sakyok hihte? 'white-footedoak', k'eweS hihte? 'custard-apple oak', and cikinib hihte? 'armadillo-eared oak'.This latter form may, for most informants, be cited in abbreviated form, i.e.simply cikinib. For some informants, the abbreviated form is, indeed, the pre-ferred usage.

Some Tzeltal speakers, however, recognize only the first three classes of oaksas (bac'il) hihte? '(genuine) oaks' and treat cikinib as being a closely related butdistinct and coordinate taxon. One Tzeltal Indian for whom the above classifica-tion of oaks holds, produced the following folk tree diagram:

hihte? 'oak'

ca?pat hihte?'excrementbarked oak'

(bac'il) hihte* 'true oak' cikinib 'armadillo-eared oak'

sakyok hihte?'white-footed

oak'

k'ewes hihte?'custard-apple

oak'

That such a situation could arise is partially explained by the fact that cikinibis by far the most divergent class of oaks referring to the native small-leafed oaksof the area (Quercus acatenangensis, Q. sapotaejolia). cikinib possesses scores of

77

LANGUAGE IN SOCIETY

objective characters which readily distinguish it from the other three broad-leafed classes of '(true) oaks'.

I lack direct evidence that the higher order taxon hihte? is at the present time afully stabilized taxon, and, as will be pointed out below, it may never become so.

In each of the examples just described it would appear that the intermediatetaxa which have arisen are each unstable as the new generics continue to be usedover time. In the case of the introduced grains, it was suggested that the newforms become conceptually 'a single unsegmented pattern', to use Geoghegan'sphrasing. In the case of ctkinib, already a monolexemic form, the prior linguisticaffiliation with hihte? is eventually lost (neutralized?) and the optional marking ofthe type generic (bac'il) hihte? will most likely be eliminated. One may predictthat the same will occur with the type generic {bac'il) ?tfim. The final result willbe the ultimate loss of the intermediate taxe, as named categories, althoughconceptually they will clearly continue to remain.

Summary of the development of intermediate names

One may summarize in tabular form the sequential steps that lead to the forma-tion and final loss of named intermediate categories in Table i.

THE LINGUISTIC RECOGNITION OF 'PLANT'

While man has no doubt tacitly recognized the world of plants as a conceptualcategory since earliest times, it does not appear to have been essential to providethe concept with a distinctive label until quite recently.

In contemporary languages of primitive peoples, a single, unique expressionfor 'plant' is notably lacking and there is no reason to assume that such was notthe case in prehistoric times. Interestingly enough, when the notion of 'plant' isexpressed, it is done via circumlocution or by the use of a form which is poly-semous with some lower order major life-form term. We might surmise anidentical situation in Theophrastus' time where it does not appear that a singlecommon expression for the full category of the plant kingdom existed. Theterm Theophrastus chose for the domain as a whole, phiton was in everydayusage the word for cultivated plant, or 'herbaceous plant of cultivation', with thesometimes restricted meaning of '(cultivated) tree' (Greene 1901: no) .

In Kirwinian and Hanunoo, the term for 'tree' can be used in some contexts torefer to plants in general (Malinowski 1933; Conklin 1954). Likewise, in Ilongot,ra?ek is polysemously 'herbaceous plant' [i.e. not 'vine' or 'tree'] and 'plant'(M. Rosaldo, personal communication). Furthermore, in Spanish, planta appearspolysemously as 'herbaceous plant, plant'; in Latin, note herba 'grass, plant',and Russian trava 'grass, plant'. These data, while not conclusive, would seem tovalidate the suspicion that the label for the unique beginner in plant taxonomiesis often drawn from one of the major class taxa, replicating a nomenclatural

78

THE GROWTH OF ETHNOBOT ANICA L NOMENCLATURE

TABLE i. Hypothetical stages involved in the formation and subsequent loss ofintermediate taxa

Path I Path II

Initial situation: nativegeneric with included

binomial specificsax bx ex

2. Introduction of a newclass not similar enoughto be included in nativepolytypic generic yetsimilar enough to allowfor concrete transposition.

3. Rise of intermediatetaxon, x', including poly-typic. native generic and xintroduced 'new' generic,.r and x' are polysemous.

bx

'foreign' x Conceptual distance of oneincluded specific is increasedbeing considered more distinct

, , *•*-*. from neighboring specifics.ax bxic (x))Such specific comes to be re-

" — " ferred to by monomial alonee.g. Tzeltal cihinib.

.v' Assumption of generic status/ \ by conceptually distinct specif-

'foreign' x x c ic, c, placing it in contrastI with typegeneric, x, and con-

,/ sequent rise of intermediateax bx taxon, x'. x and x' are poly-

semous.ax bx ex

Optional marking of type-generics with the expression

foreign x 'genuine', 'true' to contrastthem with recently formedor introduced taxa.

As the introduced item is neutralizedover time and optional marking of type-generic is eliminated leading to subse-quent loss of intermediate taxon.

foreign x

ax bx ex

Over time, prior linguistic (but notnecessarily conceptual) affiliation with xis lost (neutralized) and optional mark-ing of type-generic is eliminated leadingto subsequent loss of intermediatetaxon.

ax bx

79

LANGUAGE IN SOCIETY

process which we have seen to be quite general in other areas of ethnobotanicalnomenclature.

On the other hand, there is some evidence that the term for the plant might insome languages derive as a compound. In ancient Sumerian, the concept 'plant'was apparently designated by the conjunction of three lower-order terms whichtranslated approximately as 'tree', 'grass' and 'vegetable' (Robert McC. Adams,personal communication). This corresponds quite well with what we know ofancient Latin, where the expression of 'tree', 'herb' {arbor et herba) was used todesignate the more general concept. In this regard, Ullmann notes:

There was in Latin no generic term for 'plant' in the modern sense: arbor andherba were the most comprehensive class-concepts in the botanical field.According to a recent enquiry, the modern meaning of 'plant' is first found in

. Albertus Magnus in the 13th century, whereas the French plante did notacquire this wider sense until 300 years later (Ullmann 1963: 181).

Finally, while there is no commonly recognized term for 'plant' in Tzeltal,there are instances where something like the notion can be expressed by thecompound te?-?ak', literally 'tree-vine'. It might be suggested that we see some-thing like this going on in English folk biology, when we attempt to refer to theconcept 'living things' often by the phrase 'plants-and-animals'.

ETHNOZOOLOGICAL PARALLELS

While I have restricted my survey of the development of ethnobiologicalnomenclature to categories of plants, it should not be surprising to find ratherclose parallels in ethnozoological nomenclature. The data in this area are farfrom complete but those I have seen suggest that identical nomenclaturalprocesses are at work. Type-specific-generic name polysemy can be found inanimal names as well as in plant names. Thus, in the Chinese of Hong KongHarbor, the Karam of New Guinea, and in Guarani of Argentina, one notesexamples such as seen below (Anderson 1967:71; Dennler 1939: 233; Bulmer1968: 622).

(Chinese)hngjhal 'lobster'

Iung3hal tshatl tshoi lungjhal

'spiney lobster' 'seven-colored spiney lobster'

80


mosak 'giant rat'

mosak'giant rat'

mosak hn-ket'giant rat species'

(Karam)

Mborevi 'tapir'(Guaranf)

Mborevi Mborevi hovih'Tapirus terrestris L.' 'T. terrestris var. obscure?

Furthermore, the optional marking of the otherwise polysemous type-specific with an attributive glossed as 'genuine' is also found in animal nomencla-ture. Some selected examples are seen from Karam and Tonkawa (Bulmer1968: 624; Gatschet 1899: 160):

(Karam)yabol 'worms'

yabol (yb)'(real) worms'

yabol gwalak'large worms'

tchiixa 'field mouse*(Tonkawa)

tchiixa (dtak) tshux esaii'(genuine) field mouse' 'bogus field mouse'

Apparently, generic name - life-form name polysemy is also a process ob-served in the development of ethnozoological nomenclature. Gatschet, in hislarge Klamath grammar notes that the Klamath '. . . often use . . . wishink

81

LANGUAGE IN SOCIETY

"garter snake" for "snake", the Modocs wdm&nigh (black snake) for the sameorder of reptiles, these species being the most frequent of their kind in theirrespective countries' (1890: 11, 145). In Porno, notes collected by the zoologistC. Hart Merriam (n.d.) show the form shah as meaning polysemously 'salmon'and 'fish' while in Shoshoni of the Snake River drainage we find ?akai 'salmon,fish' (Wick Miller, personal communication). From Karam, Bulmer (1968)reports kmn as generically 'small edible mammals' but as well a life-form nameindicating all 'game animals'.

As concerns evidence indicating type-generic lexical marking analogous tothat found for plant names, Gatschet cites guato 'bird' and guato-hi 'real bird,i.e. eagle'; likewise, sane is 'snake' while sanehi refers to 'rattlesnake, i.e. realsnake' (Gatschet 1899: 157).

I have only one example of life form - unique beginner polysemy in ethno-zoological nomenclature. In Orok of Sakhalin Island, Austerlitz (1959:211)notes buju as meaning both 'bear' and 'animal'. It is not unlikely, however, thatpolysemy of this type will also be found widely as more information becomesavailable.

Finally, data from Tzeltal ethnozoology would seem to replicate in importantways some of the problems suggested in the discussion of the formation ofintermediate taxa, especially as concerns the classificatory treatment of in-troduced organisms in the case of the New Guinea Mt Hagen materials.

At the conquest, the Spanish introduced to the Tzeltal three domestic animalswhich were to assume critical cultural importance: chickens, sheep and pigs. Foreach of these introduced organisms, there existed native classes for which theforeign animals could be seen as similar, namely, birds, deer and wild pigs.

On historical linguistic grounds, we know that the Pre-conquest Tzeltalterms for these native forms were mut 'bird', cih 'deer' and citam 'wild pig'.The situation at the present time is, however, as follows:

mut 'chicken' vs. tehikil mut 'bird' (lit. 'forest chicken')cih 'sheep' vs. tehikil cih 'deer' (lit. 'forest sheep')citam 'pig' vs. wamal citam 'wild pig* (lit. 'brush pig')

In each case, the unmarked form has become restricted to the introducedanimals, while the aboriginal organisms have become linguistically marked.Full information is lacking on the complete historical cycle leading to thisoutcome, but there are sufficient data to suggest that the situation is very close tothat seen in the Mt Hagen treatment of oka.

One may note the following developmental sequence for one item of the set,cih 'deer'.

Pre-conquest:

cih 'deer'

82


Conquest:ah 'deer'

Early post conquest:(te?tikil) iih'forest deer'

Late post conquest:te?tikil ah[obligatory marking]

Present:tePttkiltih'

'sheep' [lit. 'cotton deer', at-tested in some contemporarydialects of Tzeltal]

(Optional marking of nativeform)

tunim cih

tunim ah'cotton deer'

(tunim) ah(optional marking)

ah 'sheep'[optional marking of introduced form has becomeunnecessary]

In both the Tzeltal and Mt Hagen data, it seems clear that a similar processhas been followed, leading names for introduced organisms from a lexicallymarked to lexically unmarked status, and that this process is directly related tothe high cultural significance of the organisms involved.

AN ASIDE CONCERNING NOMENCLATURAL DEVOLUTION

At the beginning, I noted that vocabularies of languages tend to increase in sizeover time. This is, of course, generally true when total vocabulary is con-sidered. Specific lexical domains, however, undergo not only growth but as welldecay. As with lexical expansion, the likely underlying causal mechanisms mustbe ultimately related to cultural evolution. Wholesale vocabulary loss in somespecific area must be due in part, at least, to the lessening of cultural importanceassociated with that particular area of human concern. Examples of such nomen-clatural devolution can be found in almost any area of vocabulary, terms foragricultural implements or carriage lexicon being obvious examples in English.

As concerns ethnobotanical nomenclature particularly, it now seems likelythat the direction of vocabulary loss will be from the particular to the general.To use the terminology of my colleagues Kay and Geoghegan, loss will occur'from the bottom up'. With little introspection, speakers of English who havebeen reared in an urban setting will recognize at once that they know virtually nospecific names for kinds of plants, that many generic names are recognizablelinguistically only as 'names' of plants, the organisms referred to being totallyunfamiliar. Nonetheless, abstract life-form names, such as 'tree' or 'grass'apparently remain as useful terms for referring to an ever-shrinking (bothliterally and figuratively) portion of one's natural environment.

83

LANGUAGE IN SOCIETY

The topic is an intriguing one deserving further study and English would beas good a language as any on which to begin research.

CONCLUSIONS

Assuming a uniformitarian view for the growth of ethnobotanical vocabulary, Ihave attempted to outline a plausible sequence of development of nomenclaturalcategories characteristic of man's linguistic recognition of the plant world. I haveargued for the primacy of generic taxa as the first ethnobiological categories tobecome encoded in a language's plant lexicon. Expansion appears to be horizontalat first, and with enlarging experience, develops both by differentiation andgeneralization. A similar argument has been made by Brown in reference tovocabulary as acquired by the child (Brown 1958) so one might suspect the pro-cess discussed here to have wider ontogenetic applicability as well.

Six major linguistic categories have been posited as sufficient for describingthe names for classes of plants in all languages. It appears highly likely thatthese categories become encoded in the history of any language in the followingorder:

flife form"! f intermediate! . .generic-*< .r >-*< . , >->unique beginner,

[specific J [varietal J

An additional claim has been made which states that with the encoding ofeach category a specifiable process of lexical marking is operable, leading plantnames from a lexically unmarked to a lexically marked status.

If the principles I have discussed prove to be general, they could allow for aplausible typological classification of the various types of ethnobotanical nomen-clature seen in the languages of the world. It is also likely that a similar classifica-tion may be appropriate for ethnozoological nomenclature. I would speculatethat the typological classification can best be interpreted as indicative of howplant nomenclature becomes encoded diachronically.

But a classification is not a theory. It is one thing to describe typologicalregularities and to suggest that their interpretation is best understood historically.It is quite a different task to outline the developmental processes involved in thechange of one system into another.

On the other hand, one usually searches for causal explanations only after onehas observed something that might be interesting to explain. Until recently,studies into the nature of the growth of vocabulary have been accorded littleimportance in linguistics and anthropology. I believe that as detailed descriptivereports reveal conclusively that aspects of man's lexicon develop in a regularlypatterned fashion, efforts towards providing theoretical explanations of theprocesses involved will be increased. While I cannot predict what mechanismswill be finally suggested as causal explanations as future work proceeds, it seems

84


likely that the details underlying the development of vocabulary must eventuallybe encompassed within some more general, technologically based theory ofcultural evolution.

R E F E R E N C E S

Almstedt, R. L. (1968). Diegueno tree: an ecological approach to a linguistic problem.IJAL 34. 9-15.

Anderson, E. N. (1967). The ethnoichthyology of the Hong Kong boat people. (Un-published Ph.D. dissertation in anthropology.) Berkeley: University of California.

Austerlitz, R. (1959). Gilyak religious terminology in the light of linguistic analysis.From the Transactions of the Asiatic Society of Japan. (Third series, vol. VII.) Tokyo.

Bartlett, H. H. (1940). History of the generic concept in botany. Bull. Torrey BotanicalClub 67. 349-62.

Berlin, B. (1969). Universal nomenclatural principles in folk science. Paper presented atthe 1968 Annual Meeting of the American Anthropological Association, New Orleans.

Berlin, B., Breedlove, D. E. & Raven, P. H. (1966). Folk taxonomies and biological clas-sification. Science 154. 273-5. Reprinted in S. Tyler (ed.) (1968). Cognitive anthropology.New York: Holt, Rinehart and Winston.

(1968). Covert categories and folk taxonomies. AmA 70. 290-9.(in press) Principles of Tzeltal plant classification. New York: Seminar Press.(n.d.) Universal principles of nomenclature and classification in folk science.

Bright, W. & Bright, J. (1965). Semantic structures in Northwestern California and theSapir-Whorf hypothesis. In E. A. Hammel (ed.), Formal semantic analysis. (AmA 67(5) pt 2.) Washington, D.C.: American Anthropological Association.

Brown, R. (1958). How shall a thing be called? Psychological Rev. 65. 14-21.Buck, C. D. (1949). A dictionary of selected synonyms in the principal Indo-European

languages. Chicago: University of Chicago Press.Bulmer, R. (1967). Why is the cassowary not a bird? A problem of zoological taxonomy

among the Karam of the New Guinea Highlands. Man 2. 1-25.(1968). Worms that croak and other mysteries of Karam natural science. Mankind 6.

621-39.(1970). Which came first, the chicken or the egghead? In J. Pouillon et P. Maranda

(eds), Echanges et communications. Melanges offerts a Claude Levi-Strauss a Voccasion deson 6oeme anniversaire. The Hague: Mouton and Co.

Bulmer, R. & Tyler, R. (1968). Karem classification of frogs. Jnl Polynesian Soc. 77.333-85.

Conklin, H. C. (1954). The relation of Hanunio culture to the plant world. (UnpublishedPh.D. dissertation in anthropology.) New Haven: Yale University.

(1962). Lexicographic treatment of folk taxonomies. In F. W. Householder and S.Saporta (eds), Problems in lexicography. (Research Center in Anthropology, Folkloreand Linguistics, Publication 21; (and) IJAL 28 (2) pt 4.) Bloomington: IndianaUniversity.

Demory, B. (1971). The word for 'tree' in the Hokan language family. In Informant.Long Beach: California State College, Department of Anthropology.

Dennler, J. G. (1939). Los nombres indigenas en Guarani de los mamiferso de la Argen-tina y paises limitrofes y su importancia para la systematica. Physis 16. 225-44.

Dentan, R. K. (1968). The Semai: a nonviolent people of Malaya. In G. Spindler andL. Spindler (eds), Case studies in cultural anthropology. New York: Holt, Rinehart andWinston.

Diamond, J. M. (1966). Classification system of primitive people. Science 151. 1102-4.Fowler, C. S. & Leland, J. (1967). Some Northern Paiute native categories. Ethnology 6.

381-404.Frake, C. O. (1962). The ethnographic study of cognitive systems. In T, Gladwin and.

LANGUAGE IN SOCIETY

W. C. Sturtevant (eds), Anthropology and human behavior. Washington, D.C.: An-thropological Society of Washington.

French, D. (i960). Types of native taxonomic process. Paper presented at the Fifty-ninthAnnual Meeting of the American Anthropological Association, Minneapolis.

Friedrich, P. (1970). Proto-Indo-European trees. Chicago: University of Chicago Press.Gatschet, A. S. (1890). The Klamath Indians of Southwestern Oregon. (Contributions to

North American Ethnology, vol. 11.) Washington, D.C.: Department of Interior, U.S.Geographical and Geological Survey of the Rocky Mountain Region.

(1899). 'Real', 'true', or 'genuine' in Indian languages. In AmA 1. 155-61. NewYork: G. P. Putnam's Sons.

Greenberg, J. H. (1966). Language universals. In T. A. Sebeok (ed.), Current trends inlinguistics, vol. 3, theoretical foundations. The Hague: Mouton and Co. 61-112.

Greene, E. L. (1909). Landmarks of botanical history. (Smithsonian Miscellaneous Col-lections, vol. 54.) Washington, D.C.

Hymes, D. (i960). More on lexicostatistics. CAnthr 1. 338-45.LeVi-Strauss, C. (1966). The savage mind. Chicago: University of Chicago Press.Malinowski, B. (1935). Coral gardens and their magic, vol. 11. New York: American Book

Company.Mathews, W. (1886). Navajo names for plants. Amer. Naturalist 20. 767-77.Merriam, C. H. (n.d.). Field checklists, Pacific Coast Region. U.S. Department of

Agriculture biological survey. (Notes in the library of the Archaeological ResearchFacility, Department of Anthropology, University of California, Berkeley.)

Paso y Trancoso, F. del (1886). La botanica entre los Nahuas. Anales del Museo Nacionalde Mexico. III. Mexico.

Roys, R. L. (1931). The ethno-botany of the Maya. (Middle American Research Series,no. 2.) New Orleans: Tulane University, Department of Middle American Research.

Stahel, G. (1944). Notes on the Arawak Indian names of plants in Surinam, jfnl N.Y.Botanical Garden 45. 268-79.

Strathern, M. (1969). Why is the Pueraria a sweet potato? Ethnology 8. 189-98.Trager, G. (1939). 'Cottonwood' = 'tree': a Southwestern linguistic trait. IJAL 9.

117-18.Ullmann, S. (1963). Semantic universals. In J. H. Greenberg (ed.), Universals of language

(2nd edition). Cambridge, Mass.: MIT Press. 217-62.Werner, H. (1954). Change of meaning: a study of semantic processes through the ex-

perimental method, jfnl General Psychology 50. 181-208.Wyman, L. C. & Harris, S. K. (1941). Navajo Indian medical ethnobotany. (Bulletin 366,

Anthropological series 3.5.) Albuquerque: University of New Mexico.

86

Date post:	13-Oct-2016
Category:	Documents
Upload:	brent
View:	225 times
Download:	3 times

Speculations on the growth of ethnobotanical nomenclature

Documents